Bonuses of the microlensing business.
Have you ever found yourself frantically searching for your car keys just before you have to go somewhere? Of course, while looking you'll find other things you either never realized you'd lost or had completely given up on.
Over the past few years, astronomers have been searching for a key too, one that may unlock a most stubborn problem in astrophysics: the nature of the dark matter that seems to pervade galaxies, galaxy clusters, and perhaps the universe as a whole. And, like you, astronomers have discovered many unexpected items during the search. This is the story of those "other" discoveries.
SEARCHING FOR THE KEY: GRAVITATIONAL MICROLENSING
Why search in the first place? In the 1930s Caltech astronomer Fritz Zwicky observed that individual galaxies within clusters move faster with respect to one another than anyone expected. In fact galaxy clusters should fly to pieces if nothing is holding them together but the gravitational attraction of the galaxies' total visible mass. Zwicky proposed, and most astronomers have since accepted, that some kind of "dark" material - stuff that does not emit detectable radiation at any wavelength - pervades the clusters, providing the mass necessary to hold them together. Astronomers believe that dark matter is present in individual galaxies and clusters of clusters as well, but they still do not know what it is. Candidates range from black holes millions of times more massive than the Sun to subatomic particles with almost no mass at all. Evidence suggests that there is more than one important type of dark matter.
About 10 years ago Bohdan Paczynski (Princeton University) proposed a way to "see" large dark objects by their gravitational effect on light from background sources. Albert Einstein's general theory of relativity predicts that massive objects should bend light rays that pass nearby. For example, a number of distant quasars appear split into two or more images as their light is deflected by the mass of an intervening galaxy or cluster of galaxies. Detecting such "gravitational lensing" at visible-light and radio wavelengths is now routine.
Paczynski realized that lensing should occur locally as well. Instead of a galaxy, the lens could be an ordinary star, but more important it could be a dark body such as a brown dwarf (failed star) or even a planet outside our solar system. If such an object passed nearly in front of a visible star in the Milky Way or a nearby galaxy, gravitational lensing would split the star's image into multiple components. Unfortunately, a brown dwarf or planet is a weak lens, so the split images would be only microarcseconds apart - utterly imperceptible with even the highest-resolution apparatus.
But that doesn't mean this process, dubbed gravitational "microlensing," is undetectable. Keep in mind that an ordinary glass lens not only changes the appearance of an object but can also make it seem larger and therefore brighter. The same occurs with a gravitational lens, and this brightening of a background object is what makes microlensing observable. Such an event would show itself as a highly unusual and distinct variation in a star's light.
Paczynski realized that random passages of dark objects in front of stars should cause well-defined, observable amplifications at a "reasonable" rate. For realistic models of our galaxy and the possible forms that lensing objects might take, he predicted that about one microlensing event should be in progress per million stars at any given time. Each event would typically last for a few days. He also suggested that microlensing would be visible even if dark objects don't exist, because ordinary stars are sufficiently abundant to act as lenses occasionally.
When Paczynski first proposed this idea in 1986, it received little reaction. No one, it seemed, was crazy enough to undertake such a huge, time-consuming project as searching for microlensing.
Soon, however, the capabilities of astronomical detectors and computers caught up with Paczynski's vision. In 1992 three groups of astronomers rose to the challenge, and within two years all three had found solid evidence that gravitational microlensing does occur. To date, more than 250 microlensing events have been recorded. Most have been found toward the constellation Sagittarius, where the vast star-swarms near the galactic center produce (and provide a fine backdrop for) microlensing events with ordinary stars as the lenses. An important subset of events has also been observed by monitoring the stars in two of our closest extragalactic neighbors, the Large and Small Magellanic Clouds.
The principal issue behind these searches - unraveling the nature of dark matter - remains unresolved. The amplitude of a given microlensing event depends on the separation of the source and lens on the sky, their relative distances, and the lens's mass. In general, we may know the light source's approximate distance, but our ignorance of the lens's location (it may be completely dark, after all) leaves its properties very poorly determined. For example, both a high-mass, nearby lens and a low-mass, distant one would generate high-amplitude, short-duration light curves that are essentially indistinguishable! The number of certain types of microlensing events and the precise rate at which they occur also remain somewhat uncertain, making it hard to understand just what the lenses are telling us.
The good news is that as the surveys continue, we'll eventually see enough events to get a statistical handle on the actual lens properties. We can anticipate that these mammoth surveys will help resolve the nature of the dark matter in our galaxy.
THE OTHER STUFF
The 250 microlensing events found so far account for about 0.000004 percent of the observations obtained. Six billion brightness measurements of nonlensed stars comprise the other 99.999996 percent. By comparison, this is 700 times as many brightness measurements as have been collected by the American Association of Variable Star Observers (AAVSO) since its founding in 1911. One of the more compelling reasons originally given to undertake these difficult surveys was to find variable stars in our own and nearby galaxies. Early on, many astronomers felt that since microlensing would be "impossible" to detect, at least the surveys would prove useful to the study of normal variables!
The best way to demonstrate the surveys' success in this regard is to consider the numbers. For example, the MACHO (MAssive Compact Halo Object) survey has detected over 40,000 variable stars in the Large Magellanic Cloud (LMC). Prior to MACHO we knew of 5,000 variables there. In one region toward the galactic center 150 variables had previously been cataloged. Our own OGLE (Optical Gravitational Lensing Experiment) project has found more than 1,000. It is as if while searching for gold we find hundreds of emeralds, rubies, and sapphires for each yellow nugget.
These surveys have led to the discovery of entirely new classes of variables as well. For example, the MACHO team has identified so-called "bumpers" - young, hot stars in the LMC experiencing surface eruptions. These outbursts have drawn immediate notice since their light curves resemble microlensing events!
Let's look at some examples that illustrate how microlensing surveys are revolutionizing our understanding of variable stars.
Suppose you struck a star with a very large hammer. What would happen? Just like a bell, the star would ring. And, just as big brass bells have different tones from big aluminum bells or small brass ones, stars vibrate most readily at certain frequencies that depend on their structure and size. But just like a bell, if you really could strike a star with a big hammer, the vibrations would fade away relatively quickly.
Well, not always. Some stars are special, and once they commence vibrating they continue indefinitely. These stars possess a delicate feedback mechanism in their atmospheres that provides a periodic "kick" just when it is needed to maintain the vibration. The energy to keep this process going comes from the heat flow out of the star's interior, like the energy supplied by someone pulling on a church bell's rope. Stars that ring indefinitely are called pulsating variables. They change in size during each pulsation period, which can range from less than an hour to a few years in different stars. Most important for astronomers, pulsating stars have trademark variations in brightness, making them easy to spot and classify.
One of the most famous types of pulsating variable is the Cepheid, named after the prototype, Delta Cephei. Cepheids are massive and therefore short-lived stars that are generally rare, but they are more common in galaxies with many young stars. The LMC is just such a galaxy, and it was already known to contain many Cepheids. Not surprisingly, microlensing surveys of the LMC found many more. The number was startling - the MACHO project alone has bagged 1,500 Cepheid variables in the central region of the LMC where only 300 were known to exist before.
These young stars usually pulsate at their so-called fundamental period, but some pulsate at shorter overtone periods. Overtones are what give musical instruments their "color" - the complex tapestry of sound that makes the same note from a trumpet, violin, and piano unmistakably different. On rare occasions both the fundamental and first overtone in a Cepheid are active simultaneously. Only 15 such double-mode (or beat) Cepheids were cataloged before the MACHO survey. Now we know of 45 in the LMC alone. The light curves of these stars are superficially similar to those of normal Cepheids except that they exhibit unusually large scatter, which indicates two concurrent periods.
The two active periods in beat Cepheids allow astronomers to precisely determine the masses of these stars independently of standard stellar evolutionary models. But getting their other detailed properties is challenging and demands extremely precise models. Until now, there haven't been enough observational data to test the models. The large number of beat Cepheids found by the MACHO survey offers a rare opportunity to dissect the interior structure of stars and to test different standard models used to derive stellar ages, distances, and other fundamental properties.
Not all pulsating stars are young. RR Lyrae stars (again, named for the prototype) are another breed, with short periods ranging from six to 18 hours. They all have similar intrinsic brightnesses, unlike Cepheids. This makes RR Lyraes very easy to use as distance indicators if you know how luminous they really are (a subject of great debate these days). They are often found in globular clusters - ancient congregations of hundreds of thousands of stars. Because these clusters are among the oldest known entities in the universe, RR Lyraes signal the presence of old stars, generally.
The various microlensing surveys have been wildly successful at finding these variables. The MACHO project has identified nearly 8,000 of them in the LMC, where only 150 were once known. Thousands of others have been found toward the galactic center, and some have been detected in the newly discovered and closest galaxy to the Milky Way - the Sagittarius dwarf.
The Sagittarius dwarf had not even been discovered when the microlensing surveys commenced. However, if other researchers hadn't done the job first, the surveys would have revealed it in short order, as it happens to lie in the direction of several survey fields. Three different projects have identified some of the many RR Lyraes that exist there. These stars are now being used to map the Sagittarius dwarf's size both across the sky (it is nearly 40 [degrees] long based on current data) and through its depth, a dimension that astronomers rarely get to study in detail.
The detection of thousands of RR Lyrae stars in the LMC is particularly impressive, suggesting that the total number of old stars there may be about twice that previously estimated. The properties of these LMC RR Lyraes also reveal that they are younger - by 2 to 4 billion years - than the oldest stars in our galaxy. Perhaps small galaxies form stars more slowly than their large cousins.
As with Cepheids, RR Lyraes can pulsate with two periods simultaneously. The MACHO study has found nearly 100 double-mode RR Lyraes, and they, too, will provide invaluable insights on the mass and internal structure of this class of stars.
STARS THAT GO BLINK IN THE NIGHT
Like people, most stars have companions. Careful studies show that binary- and even triple-star systems are very common in the Milky Way and its neighbors. Binary stars reveal themselves in many ways. Visual binaries can be resolved as a pair in a telescope, though long-term measurements may be needed to confirm whether they are physically associated. Spectroscopic binaries are revealed by their spectra, which either embody the properties of two stars at once or show the effects of orbital motion - with the spectra of one or both stars in the pair shifting periodically.
Or components of a binary can eclipse one another. Just like the Moon passing in front of the Sun, if one star moves between us and its companion, we receive less light overall. Eclipsing binaries produce very characteristic light curves, and some examples are shown at upper right. In extreme cases the stars orbit each other in only six to eight hours and are in physical contact. It would be wondrous to see such systems up dose, but we can infer much about their characteristics just from their brightness variations.
Although eclipsing binaries are not rare, the sheer number found in microlensing surveys is phenomenal. The French EROS (Experience de Recherche d'Objets Sombres) team found 79 in a very small region near the center of the LMC. The OGLE team has found 933 in a section of Baade's Window (an area in Sagittarius rich in stars and relatively free of interstellar material), where fewer than 100 had been known before. Binary stars are useful as distance indicators, and they provide valuable information about the age of the stellar population in which they were born. The binaries found in these surveys have already proved useful for studying our galaxy's structure and that of the LMC.
Binaries are so common they have even revealed themselves during microlensing events. The light path through the gravitational field of the (double) lens is far more complex, and the resulting light curve appears quite different from the case of a single lens.
MICROLENSING AND PLANETS
Binary lenses invite an interesting question: if microlensing can reveal binary stars, can it reveal planets? The very exciting recent detections of low-mass companions around nearby stars (S&T: August 1996, page 20, and May 1997, page 24) suggest that planets may be very common. If a lens - that is, a normal star - happens to have a planet, and if that planet happens to be properly aligned, it would produce a very short microlensing event. While stellar events last a few days, events due to massive planets may take only hours. Obviously, to find a planet this way you need to catch a normal lensing event in progress, then monitor the star hourly. Roughly speaking, about 10 to 100 times as many observations are needed to search for planetary microlensing as to find normal events.
Despite the added challenge, the lure of planetary detection is strong, and some of the original microlensing surveys have begun to search for planets. Other new surveys, for example, PLANET (Probing Lensing Anomalies NETwork), have been designed to follow up lensing reports in the hope of finding short-duration secondary events caused by planetary companions. The targets must be studied full-time, which means having access to a worldwide network of telescopes so that at least one or two are always under a clear night sky. The PLANET team coordinates telescopes in Australia, South Africa, and Chile to observe microlensing events found toward the galactic center. We may truly be on the verge of characterizing solar systems beyond our own, and microlensing is likely to play a big role.
LET'S MEET AT THE BAR
Although most of the emphasis in microlensing surveys has been on the study of variable stars, data collected on nonvariables have proved useful as well. Maps of the central structure of our galaxy from far-infrared satellites such as COBE suggest that the Milky Way contains a large barlike structure, similar to what is seen in many other spiral galaxies. Variable stars, in particular the RR Lyraes, are certainly useful in studying the bar's structure. But closely associated with these variables are low-mass stars that have lost matter in a red-giant phase. These so-called horizontal-branch stars all have about the same intrinsic brightness, so they can be used to estimate distances fairly reliably. The OGLE survey has monitored many fields toward the galactic center and detected large numbers of these stars. They can be confidently identified based on their brightness and color.
One intriguing feature revealed in this survey is that horizontal-branch stars found along the Milky Way northeast of the galactic center always appear brighter than the same type of star found on the opposite side of the nucleus. This suggests that one side is closer, just what you might expect for a central bar. Detailed analysis indicates that the structure is more like a fat cigar, about twice as long as wide, and oriented 20 [degrees] to 30 [degrees] away from the line connecting us and the galactic center. The diagram at upper right is a top view of our galaxy roughly showing how the bar is oriented relative to us and the nucleus.
The basic properties of the bar identified in this way agree nicely with COBE maps and other independent observations. The existence of the bar even helps explain why the rate of microlensing is higher toward the galactic center than was predicted. Our line of sight passes nearly through the bar's length, so more stars are there to act as lenses (and background sources) than expected. Further results from the surveys should soon allow us to map the three-dimensional structure of our galaxy's central parts from the distribution of microlensing effects there.
A BRIGHT OUTLOOK FOR DARK MATTER
Only four years ago microlensing was unobserved and generally believed unobservable. Now, microlensing surveys not only reveal the events they set out to detect but also have become powerful tools to study stars and how the universe changes with time. The enormous resulting databases have also helped rewrite the structure of our galaxy and trace the makeup of the Sagittarius dwarf as it plunges into our galaxy's inner halo. Like all exciting experiments, the search for microlensing has revealed far more than originally expected. Born from the desire to reveal dark matter, microlensing surveys have a very bright future.
MARIO MATEO is an astronomer at the University of Michigan and a member of the OGLE survey team. He is currently studying binary stars in globular clusters and Milky Way accretion events such as its accrual of the Sagittarius dwarf.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||data on new variable stars|
|Publication:||Sky & Telescope|
|Date:||Sep 1, 1997|
|Previous Article:||Shadows of creation: quasar absorption lines and the genesis of galaxies.|
|Next Article:||A new look at the Milky Way.|